第11講定性資料統(tǒng)計描述與推斷0604

1、定性資料的統(tǒng)計描述與推斷22表多層22表2R表多層2R表RC表非參數方法的CMH解法122表ChisqExactOR and RRMcNemar222表data bronchit;input treat$ outcome$ count;cards;jiang y 70jiang n 263dan y 102dan n 180;proc freq order=data;weight count;tables treat*outcome/chisq;run;3D3p1.prgD3p2a.prgD3p2b.prg4多層22表見CMH程序52R表見程序2R rowmeanscores 2Rrank 2

2、R scores6Row mean scoresdata cough;input treat$ response$ count;cards;test none 5 test improve 24 test notable 45 test cure 15 control none 4 control improve 23 control notable 39 control cure 18;proc freq order=data;weight count;tables treat*response/chisq CMH;RUN;7RANKdata cough;input treat$ respo

3、nse count;cards;test 1 5 test 2 24 test 3 45 test 4 15 control 1 4 control 2 23 control 3 39 control 4 18;proc freq ;tables treat*response/cmh scores=rank;weight count;RUN;proc npar1way;class treat;var response;freq count;run;8多層2R表Data cough2r;input center$;do i=1 to 2;input treat$;do j=1 to 4 ;inp

4、ut response$ count;output;end;end;cards;one testnone 7 improve 11notable 8 cure 4control none 2 improve 13notable 13 cure 2twotestnone 2 improve 7notable 22 cure 9controlnone 0 improve 6notable 20 cure 14threetestnone 5 improve 9notable 13 cure 3controlnone 6 improve 9notable 15 cure 0;run;proc freq

5、 order=data;weight count;tables center*treat*response/cmh;run;9行列表行和列變量都是名義變量行為名義變量列為順序變量行和列都是順序變量10列聯表的分類及其使用的統(tǒng)計分析方法變量的統(tǒng)計性質及其專業(yè)屬性列聯表分類統(tǒng)計分析方法XY皆為名義變量且屬性不同雙向無序檢驗，Fisher精確概率法XY之一為名義變量且屬性不同單向有序秩和檢驗，CPD，Ridit分析XY為有序變量且屬性不同雙向有序相關分析，線性趨勢檢驗XY為有序變量且屬性相同雙向有序一致性檢驗、特殊模型分析11行和列變量都是名義變量data cat5_1;input type$ di

6、sease$ count;cards;o ulcer 98 o cancer 38 o control 289a ulcer 67 a cancer 41 a control 262b ulcer 13 b cancer 8 b control 57 ab ulcer 18 ab cancer 12 ab control 30;proc freq order=data;weight count;tables type*disease/chisq nopercent norow nocol;run;12行為名義變量列為順序變量data cat5_2; input type$; do i=1 to

7、 4;input effect$ count; output; end;cards;Simple asthmanone 13 some 30 marked 18 control 65 none 18 some 36 marked 16 ontrol 77Simpemph asthemphnone 11 some 23 marked 6 control 42 none 36 some 47 marked 11 control 94;proc freq order=data;weight count;tables type*effect/chisq cmh nopercent nocol noro

8、w; run;13行列均為順序變量Data cat5_3;Input silicos density count;Cards;1 1 43 1 2 188 1 3 142 1 1 2 2 96 2 n3 723 1 6 3 2 17 3 3 55;Run;Proc corr spearman;Freq count;Var silicos density;Run;Proc freq;Weight count;Tables silicos*density/cmh nopercent norow nocol;Run;14行列為順序變量且屬性相同見程序agree15單向有序列聯表與秩和檢驗16兩樣本秩和檢驗CMH見mouse程序17完全隨機化設計資料的檢驗（H檢驗）見程序c